Neurobiology of Aging 83 (2019) 124e129

Contents lists available at ScienceDirect

Neurobiology of Aging
journal homepage: www.elsevier.com/locate/neuaging

Brain reserve, cognitive reserve, compensation, and maintenance:

operationalization, validity, and mechanisms of cognitive resilience
Yaakov Stern a, Carol A. Barnes b, Cheryl Grady c, Richard N. Jones d, Naftali Raz e, f, *
a
Cognitive Neuroscience Division, Department of Neurology, Columbia University, New York, NY, USA
b
Departments of Psychology, Neurology and Neuroscience, and Evelyn F. McKnight Brain Institute, University of Arizona, Tucson, AZ, USA
c
Rotman Research Institute, Baycrest Centre, and Departments of Psychology and Psychiatry, University of Toronto, Toronto, Ontario, Canada
d
Departments of Neurology and Psychiatry & Human Behavior, Brown University, Providence, RI, USA
e
Institute of Gerontology and Department of Psychology, Wayne State University, Detroit, MI, USA
f
Center for Lifespan Psychology, Max Planck Institute for Human Development, Berlin, Germany

a r t i c l e i n f o a b s t r a c t

Article history: Significant individual differences in the trajectories of cognitive aging and in age-related changes of brain
Received 25 February 2019 structure and function have been reported in the past half-century. In some individuals, significant
Received in revised form 5 March 2019 pathological changes in the brain are observed in conjunction with relatively well-preserved cognitive
Accepted 11 March 2019
performance. Multiple constructs have been invoked to explain this paradox of resilience, including brain
reserve, cognitive reserve, brain maintenance, and compensation. The aim of this session of the Cognitive
Aging Summit III was to examine the overlap and distinctions in definitions and measurement of these
Keywords:
constructs, to discuss their neural and behavioral correlates and to propose plausible mechanisms of
Cognitive aging
Measurement
individual cognitive resilience in the face of typical age-related neural declines.
Animal models Ó 2019 Elsevier Inc. All rights reserved.
MRI
Functional connectivity

1. Introduction them to withstand the onslaught of cognitive aging and to mitigate

Typical aging is accompanied by declines in multiple cognitive
skills, but trajectories of aging exhibit substantial individual dif-
ferences. Some individuals show precipitous deterioration, whereas
others maintain their cognitive performance to the end of life.
Although multiple factors may determine individual paths of
cognitive aging, some people appear more resilient than others to
the detrimental effects of aging and the associated pathological
changes. Although its origin remains obscure, the concept of indi-
vidual resilience received empirical support in the late 1960s, after
scientists observed a discrepancy between the burden of brain
pathology and ante-mortem cognitive performance in a series of
brain specimens (Blessed et al., 1968).
These findings, combined with later observations of recovery
from brain injury (Satz, 1993) and noninvasive neuroimaging of
patients with dementia (Stern et al., 1992), have led to the hy-
pothesis that some individuals may possess a reserve that allows
* Corresponding author at: Institute of Gerontology and Department of Psy-
chology, Wayne State University, Detroit, MI 48202, USA, and Center for Lifespan
Psychology, Max Planck Institute for Human Development, Berlin, Germany. Tel.: þ1
313 577 2297; fax: þ1 313 664 666.
E-mail address: nraz@wayne.edu (N. Raz).
the cognitive impairment expected from brain pathology. To
explain this finding, scientists have invoked multiple constructs,
including brain reserve, cognitive reserve, brain maintenance, and
compensation. Brain reserve is a neuroanatomic resource that re-
flects structural properties of the brain that somehow afford a
surplus capacity to maintain cognitive function despite substantial
loss of their material substrate. With cognitive reserve, the cogni-
tive skills and abilities that are acquired before the onset of neural
deterioration serve as a hedge againstdor even actively miti-
gatedloss of function inflicted by the progressive brain failure,
while revealing significant individual differences in resilience and
flexibility. Some people appear to be better than others in main-
taining structural and functional properties of their brains while
aging, with multiple factors contributing to variations in success.
The mechanisms by which brain and cognitive reserve bolster in-
dividual ability to maintain intact cognitive performance over time
and thus contribute to more successful cognitive aging remain a
topic for ongoing research.
The relationship between brain reserve, cognitive reserve,
maintenance, and compensation and their contribution to resil-
ience is the focus of a lively debate within the research community.
The arguments center on the core issues of definition and

0197-4580/$ e see front matter Ó 2019 Elsevier Inc. All rights reserved.
https://doi.org/10.1016/j.neurobiolaging.2019.03.022
 Y. Stern et al. / Neurobiology of Aging 83 (2019) 124e129
measurement of each construct, their differential validity, and their
neural mechanisms, as well as their potential for guiding devel-
opment of interventions to mitigate the effects of cognitive aging on
behavior. Regardless of how the related and partially overlapping
constructs are understood in the overarching notion of resilience,
they are all theoretical entities that currently cannot be measured
directly. Therefore, a problem of measuring resilience and evalu-
ating the validity of its indicators needs to be addressed.
During this session of the Cognitive Aging Summit III, chaired by
Naftali Raz, the speakers sought to define and describe these con-
structs to explain the phenomenon of resilience: brain reserve,
cognitive reserve, maintenance, and compensation. They examined
the fundamental definition and measurement issues facing the
study of resilience and discussed models and mechanisms that
describe brain adaptation to age-related changes, and, in principle,
could guide interventions to alleviate age-related cognitive
declines.
2. Reserve: an evolving concept
2.1. Yaakov Stern
The term reserve describes the difference between the degree of
brain damage observed in an individual and the clinical manifes-
tation of that damage. At least 2 models explain reserve.
The brain reserve model defines reserve as a physical trait: some
people have larger brains, with more neurons and synapses, which
may allow their brains to absorb more injury before cognitive
function is affected (Katzman et al., 1988; Satz, 1993). Brain reserve
may therefore be thought of as individual differences in the hard-
ware of the brain. Brain reserve can be conceived as a passive entity,
which reflects an overt expression of impairment only after an in-
dividual’s capacity falls below a crucial threshold of loss of the brain
substrate (Satz, 1993). People with greater initial brain capacity can
therefore tolerate more neurological attrition in absolute terms
before crossing into impairment. However, some people can
maintain their brain better than others throughout the aging pro-
cess as a function of genetics and life experienceda concept called
brain maintenance (Nyberg et al., 2012). Thus, past life experience
can influence the status of brain reserve at any time.
The cognitive reserve model posits flexibility and adaptability of
cognitive/brain networks that allow the brain to actively resist the
effects of age- or disease-related changes (Stern, 2002). The effi-
ciency and capacity of individuals’ cognitive networks may differ.
Similarly, some individuals may be better able to compensate for
brain damage by using alternate brain networks to maintain func-
tion (Cabeza, 2002; Stern, 2009). Cognitive reservedwith its focus
on functionality, plasticity, and adaptabilitydmay be thought of as
software that carries out the calculations in the brain, which is
influenced by all aspects of life experience (Stern, 2009). Thus, the
cognitive reserve model can be considered active. Given the same
amount of brain reserve, some people can better cope with age- or
pathology-related brain changes than others, depending on their
cognitive reserve capabilities.
Evidence supporting cognitive reserve includes epidemiologic
data regarding lifestyle and clinical outcomes. Individuals with
higher premorbid IQ, educational or occupational attainment, or
engagement in late-life leisure activities have a reduced risk of
developing dementia (Deary et al., 2004; Scarmeas et al., 2001;
Stern et al., 1994; Valenzuela and Sachdev, 2006) and may have
slower rates of age-related cognitive decline (Zahodne, 2015). The
underlying assumption is that some aspect of these experiences has
allowed people to better cope with age- or disease-related changes.
Thus, these experiences have often been used as proxy measures for
cognitive reserve. In addition, many studies have demonstrated
125
that, in individuals with comparable clinical severity of mild
cognitive impairment or Alzheimer’s disease, those with higher
levels of these cognitive reserve proxies have more advanced Alz-
heimer’s type pathology (Bennett et al., 2003; Ewers et al., 2013;
Stern et al., 1992). These observations indicate that individuals
with higher cognitive reserve can better cope with the brain
changes and can maintain higher function despite brain
deterioration.
Cognitive reserve is typically measured via proxies that sum-
marize experiences that affect reserve, such as educational and
occupational attainment (Stern et al., 1999). Although these mea-
sures moderate the relationship between brain changes and clinical
status, they are not direct measures of cognitive reserve and do not
reflect the underlying brain processes through which cognitive
reserve operates. Many investigators have studied the “neural
implementation” of cognitive reserve. As these relationships
become better understood, proxies for cognitive reserve may be
replaced with direct measures of specific brain processes that un-
derlie cognitive reserve. For example, our group has recently
described a pattern of brain activation that occurs when people are
performing almost any task and that is expressed to a greater de-
gree by people with higher IQ than those with lower IQ (Deary et al.,
2004). Furthermore, expression of this pattern moderates the
relationship between age-related brain changes and performance.
Another group has described a resting brain oxygenation level
dependent pattern that seems to be associated with increased
cognitive reserve (Franzmeier et al., 2017). Expression of such a
pattern may be a more direct measure of cognitive reserve, and
examination of the pattern may yield information about how
cognitive reserve is neurally implemented.
Brain reserve and cognitive reserve are not mutually exclusive.
Structural features and dynamic network capacity both play a role
in how the brain functions in the face of age-related brain changes
and pathology. Furthermore, the concept of brain maintenance
suggests that life experiences can mitigate age-related brain
changes. A better understanding of the brain’s adaptability will
enable approaches to preserve or increase both brain and cognitive
reserve to delay or reverse the effects of age-related cognitive
decline.
3. Animal models of brain adaptation and compensation in
aging
3.1. Carol A. Barnes
Although humans are unique in developing Alzheimer’s disease,
many animals display signs of neurodegeneration and cognitive
decline as they age. Studies of aging, memory, cognitive reserve,
and compensation in animals may therefore yield important in-
sights into changes in brain function during normal aging in
humans. Because it is possible to conduct much more detailed
mechanistic studies in animals than in humans, animal studies are
an important piece of the cognitive aging and decline puzzle.
With well-reasoned models, animal studies can provide
powerful insights into the molecular, cellular, and neural mecha-
nisms of human aging. To date, researchers have looked for ele-
ments of brain structure and function that are common across
species and have tried to extrapolate insights gained from these
studies to humans. Tests of spatial memorydthe ability to recall
where things are in physical or virtual three-dimensional space-
dare particularly useful when comparing data from studies of
humans and other animals because language is not an essential
aspect of that area of cognition. Spatial memory and spatial navi-
gation strategies are linked to the functional and structural prop-
erties of brain structures that are common to all mammals: the
126 Y. Stern et al. / Neurobiology of Aging 83 (2019) 124e129
hippocampus, the entorhinal cortex, and the striatum. Measure-
ments of spatial memory and spatial navigation in classic laboratory
paradigmsdfor example, the water mazedhave been extrapolated
into the human realm via virtual reality analogs (Daugherty et al.,
2016; Moffat et al., 2001) to reveal individual differences in the
associations between regional brain properties and cognition in
humans and rodents (Lester et al., 2017).
Notably, spatial memory deficits have been observed across all
animals at the ages at which they are considered to be “old,” for
example, approximately 65 years for humans, 22 years for ma-
caques, and 2 years for rats. The timing and extent of these memory
differences and changes over time, however, exhibit substantial
interindividual variability. Animal studies helped to establish that
age-related memory declines do not stem from cell death, which is
not simply a part of normal aging. In fact, within the hippocampus,
a brain region key to the formation of new memories, there is no
age-related difference in abundance of its primary cell type-
sdneither granule cells of the dentate gyrus nor CA3 or CA1 py-
ramidal cells (Gray and Barnes, 2015; Keuker et al., 2003; Rapp and
Gallagher, 1996; West, 1993).
After these observations, scientists focused the animal models of
human memory on the age-related loss of function of neurons, that
is, their plasticity, adaptability, or connectivity rather than the
number of neurons. The overarching hypothesis is that older in-
dividuals with high cognitive ability would have more plastic,
denser, or stronger connections among their neurons than would
their lower-performing peers (Barnes, 1979; Barnes and
McNaughton, 1980; Bondareff and Geinisman, 1976; Norris et al.,
1996; Rogalski et al., 2012; Stoub et al., 2012).
Aging animals frequently show loss in the components of neu-
ropil. They evidence thinner dendritic arbors and greater “axon
pruning,” a loss of pathways carrying signals from neural cell bodies
to other cells via synaptic connections. As the main components of
neuropil, axons and dendrites are essential for formation of syn-
apses and eventual arrangement of individual neurons into com-
plex circuits. Their loss during aging is therefore a plausible
explanation for impaired communication among neurons, regard-
less of the sheer number of communicating cells. Importantly,
studies in rats revealed that older rats that display fewer granule
cell synapses compared with their younger counterparts show
increased strength of the individual signals transmitted across the
remaining synapses (Barnes and McNaughton, 1980). Thus, an
adjustment in synaptic strength may form a neural substrate of
compensation, as far as it provides some degree of stability of
function across the neural circuit.
Studies in aging rodents show that although synapses onto
granule cells are stronger in older animals, the overall output from
granule cells onto CA3 neurons is still reduced. Consistent with the
idea that the brain may continuously adapt during the normative
aging process, CA3 neurons that receive less input from their
partner neurons with age are more sensitive to the input they
receive, that is, they are hyperexcitable (Foster et al., 1991; Wilson
et al., 2005). Studies in nonhuman primates also show increased
excitability of single CA3 neurons, which suggests a possible
mechanism for this change in circuit function: a decrease in the
number of inhibitory neurons in the CA3 region that contain the
neurotransmitter GABA effectively reduces levels of inhibition of
CA3 cells (Thome et al., 2016). Consistent with this observation,
magnetic resonance imaging (MRI) memory experiments in older
humans show greater activation in the CA3 region (Yassa et al.,
2011).
These studies underscore the brain’s adaptability in the face of
reduced neural resources and relative increases in external de-
mands, as outlined in a recent model of brain adaptability and
plasticity (Lӧvdén et al., 2010). Elucidation of the molecular and
cellular mechanisms that control synapse numbers and function
throughout the life span may provide important insights into the
key mechanisms of cognitive decline. To advance this goal, we must
improve translational strategies that bridge animal and human
research on aging.
4. Age differences in the dynamic flexibility of brain activity
and functional connectivity
4.1. Cheryl L. Grady
The dynamic flexibility of the brain allows it to incorporate and
react to new information in real time and to rebalance or form new
neural connections to adapt to new circumstances (Garrett et al.,
2013). In older adults (e.g., at least 65 years of age), several neural
processes that are fundamentally associated with dynamic flexi-
bility are altered. These include differentiation, defined as stimulus-
related or task-related specificity of activity in various neural net-
works, and functional connectivity among brain networks.
Studies of task dynamics in older adults show dedifferentiation
of brain activity during memory and other tasks (i.e., reduced
selectivity). This observation has led to the hypothesis that selective
brain activity is important for some subset of memory tasks; it
further suggests that reduced selectivity may be a mechanism un-
derlying the observed age differences. For example, a representa-
tive memory function that shows age-related differences is
associative memory, which reflects the ability to bind together
disparate pieces of information and to form a unified representa-
tion of the combined stimulus (Old and Naveh-Benjamin, 2008). In
a typical experiment to test this memory skill, participants are
shown pairs of unrelated images and later are asked to discriminate
represented (“old”) pairs from new pairs that the participants have
not seen. On average, older adults remember these paired stimuli
more poorly than do younger adults. When the tests are adminis-
tered as a part of a functional MRI (fMRI) experiment, older par-
ticipants show reduced selectivity in brain regions that are
specifically involved in encoding the paired images (Saverino et al.,
2016), consistent with their reduced associative memory. This
provides evidence that reduced selectivity (dedifferentiation) of
brain activity during encoding, one type of dynamic brain activity,
may be an important mechanism underlying reduced memory
performance in older adults.
The past 2 decades saw a surge in studies of task-free (resting
state) connection dynamics of brain activity in fMRI experiments,
providing another measure of brain flexibility, that is, the functional
connectivity within and between brain networks at rest. In this
paradigm, time-dependent fMRI signals recorded in specific
neuroanatomic locations are examined for coherent and covarying
activity, with higher coherence signifying stronger functional con-
nectivity. Three functional brain networks of particular interest are
the default mode network, the dorsal attention network, and the
frontoparietal control network, which have been linked to internal
attention, external attention, and cognitive control, respectively
(Spreng et al., 2013). In general, older adults have weaker connec-
tions within nodes of an individual network but stronger connec-
tions between different networks (Chan et al., 2014). Frontoparietal
control network integration is inversely related to default mode
network connectivity in older adults, with stronger between-
network connections in the frontoparietal control network in
those individuals with weaker connections within the default mode
network (Grady et al., 2016). In addition, frontoparietal control
network integration is positively correlated with memory perfor-
mance in older adults. These observations raise that possibility that
more robust frontoparietal control network connectivity might be a
source of cognitive reserve.
 Y. Stern et al. / Neurobiology of Aging 83 (2019) 124e129
Both lines of research converge on the notion of robust age-
related and individual differences in the dynamic range of brain
activity. Age-related reduction in selectivity of task-related brain
activity is linked to poorer associative memory, whereas increased
integration of the frontoparietal control network at rest is associ-
ated with better memory performance in older adults. Importantly,
dynamic brain activity is related to cognitive performance regard-
less of age and therefore elucidating this aspect of brain function
may advance understanding of cognition in general as well as age-
related differences therein.
5. (Mis-) measurement of reserve
5.1. Richard N. Jones
The brain pathology that is often seen in demented individuals
can also be present in cognitively normal elderly adults who show
no or very limited signs of cognitive decline (Blessed, 1968; Satz,
1993). This observation has led to the hypothesis that some
elderly adults may possess a reserve that allows them to continue to
function well in the face of sometimes significant brain pathology.
Although many researchers accept the concept of brain or cognitive
reserve, it has proven difficult to define it specifically and to
quantify it accurately. Several biological traits (e.g., brain size) (Satz,
1993) and socialdemographic characteristics (e.g., years of educa-
tion) (Stern et al., 1999) have been associated with reserve. Yet, it
remains unclear whether and how these factors are fundamental to
reserve as a construct.
The uncertainty surrounding the exact nature of reserve has led
to approaches to measure reserve as a latent, or hidden, variable
(Jones et al., 2011). Latent variables represent theoretical constructs
that cannot be directly observed and measured. Their existence can
be inferred through observation and modeling of other variables
that can be directly observed and measured, that is, indicators. The
uncertainty of the relationship between the construct and its in-
dicators is a major theoretical issue in research on reserve and
resilience and it hinges on the distinction between the formative
and reflective natures of the postulated entity. In the formative
model, the latent variable is created by and is a consequence of the
measurable variables. In the reflective model, the measurable var-
iables are viewed as being caused by and reflecting one or more
facets of the underlying, unseen but theoretically derived, latent
variable. In either case, carefully identifying and reliably measuring
the appropriate indicators is key because they provide insights into
the unknown, latent factor.
The distinction between formative and reflective constructs is
not just a matter of classification. These 2 models lead to distinct
interpretation of the empirical data and have very different impli-
cations for designing interventions to mitigate age-related cogni-
tive decline. If the formative model is correct, then measured
variablesdsuch as educationdcause the accumulation of reserve,
and increasing years of education would make an individual less
susceptible to cognitive decline. If the reflective model is correct,
then some yet-unknown factor influences education, reserve, and a
host of other characteristics, and identifying appropriate reflective
indicators is key. These theory-derived indicators would need to be
tested in observational or experimental settings to determine
whether they adequately account for lifetime cognitive ability, with
formative versus reflective models explicitly compared in their fit to
the data (Hagger-Johnson et al., 2011).
Reserve, as it is currently conceived, is an important modulating
factor between brain decline (gradual or catastrophic) and cogni-
tive outcomes. To understand and measure the impact of reserve on
this relationship, it is important that researchers accurately assess
the behavioral and cognitive indicators as well as accurately
127
measure normal and pathological changes in brain structure and
function. Measures of cognitive function are vulnerable to con-
founding variables, including possible selection bias in samples, the
effect of practice, and mere exposure to multiple testing, as well as
artifactual effects related to floors and ceilings in cognitive perfor-
mance. Defining the influence on cognitive aging of these variables
and devising methods to measure them are important steps to
improve our understanding of aging. In summary, fine-tuning
theoretical concepts and developing reliable and valid indicators
should go together in promoting our understanding of resilience
and harnessing its power for improving the cognitive trajectories of
older adults.
6. Discussion
No single approach will fully reveal how cognitive resilience is
implemented in the brain. The emphasis should be on multimodal
neuroimaging studies that will combine noninvasive neuroimaging
methods geared to simultaneously evaluate neuropil volume,
myelin content, white-matter connectivity, and axonal integrity, as
well as functional organization and key neurotransmitters over
multiple occasions throughout life span. More studies should
address the neuroenergetic (Raz and Daugherty, 2018) and vascular
(Raz and Rodrique, 2006) correlates of age-related changes in brain
structure and function.
The limitations of cross-sectional designs and their inability to
capture the dynamics of aging even in a single modality, not to
mention the brain-cognitive relationship, has been highlighted by
many methodologists (Hofer and Sliwinski, 2001; Lindenberger
et al., 2011; Maxwell and Cole, 2007). Overcoming this limitation
is a daunting task, and it is made even more challenging by the fact
that researchers’ life spans are commensurable with that of the
subjects of their study. Thus, there is a need for long-term longi-
tudinal studies with multiple measurement occasions. Moreover,
refinement of the existing animal models and development of new
ones that optimize the demands for manageable life span, sufficient
neuroanatomical and neurofunctional homology, and comparable
pattern of age-related diseases and risk factors is a critical need.
The challenges notwithstanding animal models are of critical
importance. Only in well-controlled animal studies can we examine
synaptic and cellular indicators of resilience with a full range of
invasive techniques. To optimize this path of investigation, it is
necessary to take a translational harmonization approach. That is,
traditional techniques of animal research (e.g., intracellular re-
cordings, genetic and epigenetic manipulations, radiolabeling)
should be supplemented with noninvasive neuroimaging that
replicates in the closest possible manner human methods of brain
mapping. Existing studies of this type have already yielded
important insights into the nature of regional volume changes
observed on MRI in humans; for example, a recent study of rodents
demonstrated the parallelism between stress-induced changes in
dendritic arborization and MRI-derived volumes of the hippocam-
pus and anterior cingulate gyrus (Kassem et al., 2013).
The challenges posed by the extraordinary complexity of the
brain can be ameliorated by adopting the mathematical methods
that have been developed in other disciplines that study complex
systems (Gu et al., 2015). Adopting the framework of control theory
to analyze age-related differences and changes in configuration of
brain networks may help elucidate the mechanisms of resilience
through understanding of preservation and maintenance of
fundamental connectivity as well as (potentially compensatory)
response of the aging brain to structural deterioration in the
context of cognitive reserve.
Because resilience is by all accounts built over a lifetime, it is
important to address its developmental roots and the risk factors
128 Y. Stern et al. / Neurobiology of Aging 83 (2019) 124e129
that modify its expression over time. Thus, life span longitudinal
studies are essential for elucidating how brain reserve emerges,
how cognitive reserve is built, how brain maintenance is imple-
mented, and how compensation is applied. In that context, life span
changes in plasticity and the ability to meet systemic and envi-
ronmental challenges that come with aging are particularly
important (Lӧvdén et al., 2010). Attention should be paid to mul-
tiple factors shaping the trajectories of brain aging and potentially
acting as modifiers of resilience, such as physical fitness (Hillman
et al., 2008); social engagement; as well as vascular (Raz and
Rodrique, 2006), metabolic, and neuroinflammatory risk factors,
including risk-associated genetic variance (Raz et al., 2015); as well
as socioeconomic deprivation (Johnson et al., 2013) with associated
high levels of stress (McEwen, 2013); and environmental pollution
exposure (Woodward et al., 2017).
Whatever the source of resilience, while it is associated with a
reduced risk of developing dementia, it isdperhaps para-
doxicallydalso associated with a faster rate of cognitive decline and
mortality once Alzheimer’s disease is diagnosed (Stern et al., 1999).
This rapid decline may occur because in Alzheimer’s disease, the
brains of more highly educated people reach relatively greater
functional and structural damage than those of less-educated in-
dividuals before the cognitive expression of the disease becomes
noticeable. This discrepancy between brain state and cognitive
performance suggests superior maintenance and, possibly
compensation, mechanisms that are being stretched to the
breaking point before they fail.
One of the central challenges is the difficulty in clearly defining
the terms and confronting the fallacy of concreteness: the belief
that something exists only because it can be measured. As a cor-
ollary, just because something can be measured does not mean that
it is the most important thing to measure. If reserve is a real thing,
then it should manifest itself in multiple contexts, if researchers can
find ways to detect it in different circumstances. Multiple sightings
of reserve in action make its existence highly plausible and inspire
continued efforts to improve understanding of this potential key
factor in ameliorating age-related cognitive impairment.
Disclosure
The authors report no conflict of interest.
Acknowledgements
The authorship order reflects the order of presentations in the
session. The Cognitive Aging Summit III was supported by a grant
from McKnight Brain Research Foundation to the Foundation for
the National Institutes of Health. The authors gratefully acknowl-
edge our sources of funding: National Institutes of Health grant R01
AG026158 to YS; National Institutes of Health grants R01
AG003376, R01 AG050548, and McKnight Brain Research Founda-
tion grant to CAB; Canadian Institutes of Health Research grant
MOP14036, the Canada Research Chairs program, the Ontario
Research Fund, the Canadian Foundation for Innovation, and the
Heart and Stroke Foundation Centre for Stroke Recovery to CG;
National Institutes of Health grant P01 AG031720 to RNJ; and Na-
tional Institutes of Health grant R01 AG011230 to NR. David Miller,
Samuel Thomas, and Nancy Tuvesson of Rose Li and Associates, Inc.
provided writing and editorial support for the preparation of this
manuscript.
References
Barnes, C.A., 1979. Memory deficits associated with senescence: a neurophysio-
logical and behavioral study in the rat. J. Comp. Physiol. Psychol. 93, 74e104.
Barnes, C.A., McNaughton, B.L., 1980. Physiological compensation for loss of afferent
synapses in rat hippocampal granule cells during senescence. J. Physiol. 309,
473e485.
Bennett, D.A., Wilson, R.S., Schneider, J.A., Evans, D.A., Mendes de Leon, C.F.,
Arnold, S.E., Barnes, L.L., Bienias, J.L., 2003. Education modifies the relation of AD
pathology to level of cognitive function in older persons. Neurology 60,
1909e1915.
Blessed, G., Tomlinson, B.E., Roth, M., 1968. The association between quantitative
measures of dementia and of senile change in the cerebral grey matter of
elderly subjects. Br. J. Psychiatry 114, 797e811.
Bondareff, W., Geinisman, Y., 1976. Loss of synapses in the dentate gyrus of the
senescent rat. Am. J. Anat. 145, 129e136.
Cabeza, R., 2002. Hemispheric asymmetry reduction in older adults: the HAROLD
model. Psychol. Aging 17, 85e100.
Chan, M.Y., Park, D.C., Savalia, N.K., Petersen, S.E., Wig, G.S., 2014. Decreased
segregation of brain systems across the healthy adult lifespan. Proc. Natl. Acad.
Sci. U. S. A. 111, E4997eE5006.
Daugherty, A.M., Bender, A.R., Yuan, P., Raz, N., 2016. Changes in search path
complexity and length during learning of a virtual water maze: age differences
and differential associations with hippocampal subfield volumes. Cereb. Cortex
26, 2391e2401.
Deary, I.J., Whiteman, M.C., Starr, J.M., Whalley, L.J., Fox, H.C., 2004. The impact of
childhood intelligence on later life: following up the Scottish mental surveys of
1932 and 1947. J. Pers. Soc. Psychol. 86, 130.
Ewers, M., Insel, P.S., Stern, Y., Weiner, M.W., 2013. Cognitive reserve associated with
FDG-PET in preclinical Alzheimer disease. Neurology 80, 1194e1201.
Foster, T.C., Barnes, C.A., Rao, G., McNaughton, B.L., 1991. Increase in perforant path
quantal size in aged F-344 rats. Neurobiol. Aging 12, 441e448.
Franzmeier, N., Duering, M., Weiner, M., Dichgans, M., Ewers, M., 2017. Alzheimer’s
Disease Neuroimaging I. Left frontal cortex connectivity underlies cognitive
reserve in prodromal Alzheimer disease. Neurology 88, 1054e1061.
Garrett, D.D., Samanez-Larkin, G.R., MacDonald, S.W., Lindenberger, U.,
McIntosh, A.R., Grady, C.L., 2013. Moment-to-moment brain signal variability:
a next frontier in human brain mapping? Neurosci. Biobehav. Rev. 37,
610e624.
Grady, C., Sarraf, S., Saverino, C., Campbell, K., 2016. Age differences in the functional
interactions among the default, frontoparietal control, and dorsal attention
networks. Neurobiol. Aging 41, 159e172.
Gray, D.T., Barnes, C.A., 2015. Distinguishing adaptive plasticity from vulnerability in
the aging hippocampus. Neuroscience 309, 17e28.
Gu, S., Pasqualetti, F., Cieslak, M., Telesford, Q.K., Yu, A.B., Kahn, A.E., Medaglia, J.D.,
Vettel, J.M., Miller, M.B., Grafton, S.T., Bassett, D.S., 2015. Controllability of
structural brain networks. Nat. Commun. 6, 8414.
Hagger-Johnson, G., Batty, G.D., Deary, I.J., von Stumm, S., 2011. Childhood socio-
economic status and adult health: comparing formative and reflective models
in the Aberdeen Children of the 1950s Study (prospective cohort study).
J. Epidemiol. Community Health 65, 1024e1029.
Hillman, C.H., Erickson, K.I., Kramer, A.F., 2008. Be smart, exercise your heart: ex-
ercise effects on brain and cognition. Nat. Rev. Neurosci. 9, 58e65.
Hofer, S.M., Sliwinski, M.J., 2001. Understanding ageing: an evaluation of research
designs for assessing the interdependence of ageing-related changes. Geron-
tology 47, 341e352.
Johnson, N.F., Kim, C., Gold, B.T., 2013. Socioeconomic status is positively correlated
with frontal white matter integrity in aging. Age (Dordr.) 35, 2045e2056.
Jones, R.N., Manly, J., Glymour, M.M., Rentz, D.M., Jefferson, A.L., Stern, Y., 2011.
Conceptual and measurement challenges in research on cognitive reserve. J. Int.
Neuropsychol. Soc. 17, 593e601.
Kassem, M.S., Lagopoulos, J., Stait-Gardner, T., Price, W.S., Chohan, T.W., Arnold, J.C.,
Hatton, S.N., Bennett, M.R., 2013. Stress-induced grey matter loss determined by
MRI is primarily due to loss of dendrites and their synapses. Mol. Neurobiol. 47,
645e661.
Katzman, R., Terry, R., DeTeresa, R., Brown, T., Davies, P., Fuld, P., Renbing, X., Peck, A.,
1988. Clinical, pathological, and neurochemical changes in dementia: a sub-
group with preserved mental status and numerous neocortical plaques. Ann.
Neurol. 23, 138e144.
Keuker, J.I.H., Luiten, P.G.M., Fuchs, E., 2003. Preservation of hippocampal neuron
numbers in aged rhesus monkeys. Neurobiol. Aging 4, 157e165.
Lester, A.W., Moffat, S.D., Wiener, J.M., Barnes, C.A., Wolbers, T., 2017. The aging
navigational system. Neuron 9, 1019e1035.
Lindenberger, U., von Oertzen, T., Ghisletta, P., Hertzog, C., 2011. Cross-sectional age
variance extraction: what’s change got to do with it? Psychol. Aging 26, 34e47.
Lövdén, M., Bäckman, L., Lindenberger, U., Schaefer, S., Schmiedek, F., 2010.
A theoretical framework for the study of adult cognitive plasticity. Psychol. Bull
136, 659e676.
Maxwell, S.E., Cole, D.A., 2007. Bias in cross-sectional analyses of longitudinal
mediation. Psychol. Methods 12, 23e44.
McEwen, B.S., 2013. The brain on stress: toward an integrative approach to brain,
body, and behavior. Perspect. Psychol. Sci. 8, 673e675.
Moffat, S.D., Zonderman, A.B., Resnick, S.M., 2001. Age differences in spatial memory
in a virtual environment navigation task. Neurobiol. Aging 22, 787e796.
Norris, C.M., Korol, D.L., Foster, T.C., 1996. Increased susceptibility to induction of
long-term depression and long-term potentiation reversal during aging.
J. Neurosci. 16, 5382e5392.
Nyberg, L., Lövdén, M., Riklund, K., Lindenberger, U., Bäckman, L., 2012. Memory
aging and brain maintenance. Trends Cogn. Sci. 16, 292e305.
 Y. Stern et al. / Neurobiology of Aging 83 (2019) 124e129
Old, S.R., Naveh-Benjamin, M., 2008. Differential effects of age on item and asso-
ciative measures of memory: a meta-analysis. Psychol. Aging 23, 104e118.
Rapp, P.R., Gallagher, M., 1996. Preserved neuron number in the hippocampus of
aged rats with spatial learning deficits. Proc. Natl. Acad. Sci. U. S. A. 93,
9926e9930.
Raz, N., Daugherty, A.M., 2018. Pathways to brain aging and their modifiers: free-
radical induced energetic and neural decline in senescence (FRIENDS) model.
Gerontology 64, 49e57.
Raz, N., Rodrigue, K.M., 2006. Differential aging of the brain: patterns, cognitive
correlates and modifiers. Neurosci. Biobehav. Rev. 30, 730e748.
Raz, N., Daugherty, A.M., Bender, A.R., Dahle, C.L., Land, S., 2015. Volume of the
hippocampal subfields in healthy adults: differential associations with age and a
pro-inflammatory genetic variant. Brain Struct. Funct. 220, 2663e2674.
Rogalski, E., Stebbins, G.T., Barnes, C.A., Murphy, C.M., Stoub, T.R., George, S.,
Ferrari, C., Shah, R.C., deToledo-Morrell, L., 2012. Age-related changes in para-
hippocampal white matter integrity: a diffusion tensor imaging study. Neuro-
psychologia 50, 1759e1765.
Satz, P., 1993. Brain reserve capacity on symptom onset after brain injury: a
formulation and review of evidence for threshold theory. Neuropsychology 7,
273e295.
Saverino, C., Fatima, Z., Sarraf, S., Oder, A., Strother, S.C., Grady, C.L., 2016. The
associative memory deficit in aging is related to reduced selectivity of brain
activity during encoding. J. Cogn. Neurosci. 28, 1331e1344.
Scarmeas, N., Levy, G., Tang, M.X., Manly, J., Stern, Y., 2001. Influence of leisure ac-
tivity on the incidence of Alzheimer’s disease. Neurology 57, 2236e2242.
Spreng, R.N., Sepulcre, J., Turner, G.R., Stevens, W.D., Schacter, D.L., 2013. Intrinsic
architecture underlying the relations among the default, dorsal attention, and
frontoparietal control networks of the human brain. J. Cogn. Neurosci. 25,
74e86.
Stern, Y., 2002. What is cognitive reserve? Theory and research application of the
reserve concept. J. Int. Neuropsychol. Soc. 8, 448e460.
Stern, Y., 2009. Cognitive reserve. Neuropsychologia 47, 2015e2028.
129
Stern, Y., Alexander, G.E., Prohovnik, I., Mayeux, R., 1992. Inverse relationship be-
tween education and parietotemporal perfusion deficit in Alzheimer’s disease.
Ann. Neurol. 32, 371e375.
Stern, Y., Gurland, B., Tatemichi, T.K., Tang, M.X., Wilder, D., Mayeux, R., 1994. In-
fluence of education and occupation on the incidence of Alzheimer’s disease.
J. Am. Med. Assoc. 271, 1004e1010.
Stern, Y., Albert, S., Tang, M.X., Tsai, W.Y., 1999. Rate of memory decline in AD is
related to education and occupation: cognitive reserve? Neurology 53,
1942e1957.
Stoub, T.R., Barnes, C.A., Shah, R.C., Stebbins, G.T., Ferrari, C., deToledo-Morrell, L.,
2012. Age-related changes in the mesial temporal lobe: the parahippocampal
white matter region. Neurobiol. Aging 33, 1168e1176.
Thome, A., Gray, D.T., Erickson, C.A., Lipa, P., Barnes, C.A., 2016. Memory impairment
in aged primates is associated with region-specific network dysfunction. Mol.
Psychiatry 21, 1257e1262.
Valenzuela, M.J., Sachdev, P., 2006. Brain reserve and dementia: a systematic review.
Psychol. Med. 36, 441e454.
West, M.J., 1993. Regionally specific loss of neurons in the aging human hippo-
campus. Neurobiol. Aging 14, 287e293.
Wilson, I.A., Ikonen, S., Gallagher, M., Eichenbaum, H., Tanila, H., 2005. Age-asso-
ciated alterations of hippocampal place cells are subregion specific. J. Neurosci.
25, 6877e6886.
Woodward, N.C., Pakbin, P., Saffari, A., Shirmohammadi, F., Haghani, A., Sioutas, C.,
Cacciottolo, M., Morgan, T.E., Finch, C.E., 2017. Traffic-related air pollution
impact on mouse brain accelerates myelin and neuritic aging changes with
specificity for CA1 neurons. Neurobiol. Aging 53, 48e58.
Yassa, M.A., Lacy, J.W., Stark, S.M., Albert, M.S., Gallagher, M., Stark, C.E.L., 2011.
Pattern separation deficits associated with increased hippocampal CA3 and
dentate gyrus activity in nondemented older adults. Hippocampus 21, 968e979.
Zahodne, L.B., Stern, Y., Manly, J.J., 2015. Differing effects of education on cognitive
decline in diverse elders with low versus high educational attainment. Neuro-
psychology 29, 649e657.